Download the Most Up-to-date Project Summary

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
HealthCyberMap: Mapping the Health Cyberspace Using
Hypermedia GIS and Clinical Codes
PhD Research Project
Maged Nabih Kamel Boulos
MBBCh, MSc (Derm & Vener), MSc (Medical Informatics)
1 Summary
“The application of GIS is limited only by the imagination of those who use it.”
—Jack Dangermond, President, ESRI
HealthCyberMap (<http://healthcybermap.semanticweb.org>) is a Semantic Web service
for healthcare professionals and librarians, patients and the public in general that aims at
mapping parts of medical/ health information resources in cyberspace in novel ways to
improve
their
retrieval
and
navigation.
The
Semantic
Web
(<http://www.w3.org/2001/sw/> and <http://www.semanticweb.org>) aims to be the
next-generation World Wide Web by giving machine-readable semantics and context to
the currently presentation-based Web pages.
HealthCyberMap features an unconventional use of GIS (Geographic Information
Systems) to map conceptual spaces occupied by collections of medical/ health
information resources. Besides mapping the semantic and non-geographical aspects of
these resources using suitable spatial metaphors, HealthCyberMap also collects and maps
the geographical provenance of these resources. Some of HealthCyberMap Web
interfaces are visual (maps for browsing resources by clinical/ health topic, by provenance
and by type), while others are textual (multilingual interfaces for browsing resources by
language, and a directory of topical resource categories, besides HealthCyberMap
Semantic Subject Search Engine that goes beyond conventional free-text and keywordbased search engines, and supports synonyms, disease variants, subtypes, as well as some
semantic relationships between terms).
HealthCyberMap adopts a clinical metadata framework built upon a clinical coding
scheme (vocabulary or ontology—ICD-9-CM clinical classification in the current pilot
service). Clinical coding schemes serve as a reliable common backbone for topical
resource indexing, automated topical classification, topical visualisation and navigation of
coded resource pools (using suitable metaphors), and enhanced information retrieval and
linking. A resource metadata base based on Dublin Core metadata set with
HealthCyberMap’s own extensions holds information about selected high-quality
resources. HealthCyberMap then uses GIS spatialisation methods to generate interactive
navigational cybermaps from the metadata base. These visual cybermaps are based on
familiar metaphors for image-word association to give users a broad overview and
understanding of what is available in this complex conceptual space of medical/ health
Internet resources and help them navigate it more efficiently and effectively.
HealthCyberMap cybermaps can be considered as semantically-spatialised, ontologybased browsing views of the underlying resource metadata base. Using a clinical coding
scheme as a metric for spatialisation (“semantic distance”) is unique to HealthCyberMap
and is very much suited for the semantic categorisation and navigation of medical/ health
Internet information resources. HealthCyberMap also introduces a useful form of
cyberspatial analysis for the detection of topical coverage gaps in its resource pool using
choropleth (shaded) maps of human body systems. The project features a cost-effective
method for serving Web hypermaps with dynamic metadata base drill-down functionality.
It also demonstrates the feasibility of Electronic Patient Record to Online Information
Services (like HealthCyberMap) Problem to Knowledge Linking using clinical codes as
crisp problem-knowledge linkers or knowledge hooks.
The Semantic Subject Search Engine queries the same HealthCyberMap resource
metadata base. Explicit concepts in resource metadata map onto a brokering domain
ontology (ICD-9-CM) allowing the search engine to infer implicit meanings (synonyms
and semantic relationships) not directly mentioned in either the resource or its metadata.
Similarly, user queries would map to the same ontology allowing the search engine to
infer the implicit semantics of user queries and use them to optimise retrieval.
A formative evaluation study of HealthCyberMap pilot service using an online user
evaluation questionnaire, in addition to analysis of HealthCyberMap server transaction
log, has been conducted during the period from 18 April 2002 to 1 June 2002 with very
encouraging results. This two-method evaluation approach was guided by methodologies
described in NIH† Web Site Evaluation and Performance Measures Toolkit among other
resources.
Many exciting future possibilities have been also investigated by the author, including the
further development HealthCyberMap as a customisable, location-based medical/ health
information service.
Keywords
Semantic Web, Resource Description Framework (RDF), Metadata, Dublin Core, Ontologies,
Controlled Vocabularies, Clinical Codes, Problem to Knowledge Linking, Clinical Decision
Support, Geographic Information Systems (GIS), Cartography, Cyberspace, Cybergeography,
Cyberspatial Analysis, Medical/ Health Information Spaces, Spatialisation, Spatial/ Visual
†
National Institutes of Health, US
Metaphors, Internet/ Web Resource Visualisation, Web Browsing/ Navigation, Cybermaps,
Web Maps, Hypermaps, Hypermedia, Digital Libraries, Information Retrieval, Medical
Search Engines, Location-based Services, Evaluation of Online Information Services, Server
Transaction Log Analysis
2 Aims and Objectives
Background and Motivation
The Internet is a system with unimaginable complexity. It has been said the Web is its
own map. But as we surf the Web, we can only appreciate a very small part of it at
any given time. We cannot for example figure out the relations of the page/ site we
are visiting to the rest of the Web, how it measures compared to the rest of the Web,
or how the rest of the Web looks like. We can have difficulties finding our way back
to a resource we visited just few hours ago. We cannot easily locate medical/ health
Web resources covering narrower, broader, related or similar topics as the one we are
currently looking at, and except when we have no particular goal(s) in mind, we
cannot plan our Web journeys ahead. We need a map or set of maps for this purpose.
Ideally, we should be also able to link our health-related and clinical questions to such
maps, and use them (or the underlying information) to find the appropriate answers to
these questions in the right place and at the right time.
Maps are well known as powerful graphic tools that can be used to classify, visualise,
communicate and navigate spatial and/ or spatialised relations (relations projected into
some conceptual space for mapping) in worlds that are too large and too complex to
be seen directly. Maps build on humans’ powerful spatio-cognitive abilities. Good
Web navigational maps should be also based on sound cartographic principles and
variables (to ensure their spatial legibility and utility), and adapted to the unique
nature of the Web.
GIS (Geographic Information Systems) take simple cartography one step further by
providing contextual links between maps and underlying databases (where attributes
of features on the maps are stored). On the Web these links can be implemented as
sensitive clickable maps (hypermaps). The quality and utility of such maps will
obviously depend on the quality of the information in the underlying databases.
Aims and Mission Statement
HealthCyberMap (<http://healthcybermap.semanticweb.org>) is a Web-based service
that aims at mapping medical/ health information resources in cyberspace in unique
and novel ways to deliver a semantically superior experience to resource providers
and consumers (the general public, patients, healthcare professionals and librarians),
helping them plan, manage, search and navigate this complex “virtual” space more
efficiently and effectively. This is achieved through semantic indexing, “intelligent”
categorisation, and interactive hypermedia visualisation of the medical/ health
information cyberspace using metadata, clinical codes and GIS technologies.
HealthCyberMap only attempts to map parts of available health and healthcare
Internet resources and services.
Objectives
(with measurable outcomes)
A navigational hypermap for browsing collections of Internet information resources
should be ideally driven by an underlying metadata base that stores meta-information
(information about information) describing these resources. HealthCyberMap
comprises two main arms or layers. The top-level visualisation/ navigation arm
(interface layer) is founded upon a robust semantic layer.
Semantic Arm Objectives
1. To define and model a suitable framework that tailors existing generic metadata
standards/ recommendations to suit the description of medical/ health information
resources.
2. To improve resource metadata quality and semantics through proper use of
clinical codes from a controlled vocabulary as descriptors of medical/ health
resource contents.
3. To propose and examine ways of reasoning semantically with resource metadata
(including clinical codes) beyond simple queries, and combining metadata with
other related and useful knowledge bases (ontologies or vocabularies) for the
purpose of customising the language, interface and content of material delivered
to consumers and offering them location-specific information.
4. To build a pilot metadata base describing a small part of medical/ health Internet
resources based on the above framework, with the application of suitable resource
quality benchmarking measures.
5. To develop a pilot Medical Semantic Subject Search Engine that outperforms
conventional free-text and keyword-based search engines, and supports synonyms,
disease variants, subtypes, as well as some semantic relationships between
medical/ health terms.
6. To develop a pilot clinical Problem to Knowledge Linking solution based on
clinical codes to contextually link the Electronic Patient Record to the resource
metadata base.
Visualisation/ Navigation Arm Objectives
7. To find an effective semantic metric based on clinical codes that can serve as a
basis for the topical mapping of medical/ health information resources in
cyberspace.
8. To find and use suitable spatialisation methods and familiar metaphors to visually
browse and navigate medical/ health information resources on the Internet in
different (but complementary) ways based on clinical codes and other metadata
elements in the resource metadata base. This objective will be complemented by
carrying a review/ critique of some existing and related cybermapping projects.
9. To use GIS to automate the classification and generation of spatialised browsing
views (navigational cybermaps) of the resource metadata base based on the above
cybermapping methods.
10. To develop a method for identifying topical coverage gaps (infogaps) in mapped
resources.
11. To develop a cost-effective solution for serving hypermaps with dynamic database
links/ drill-down functionality on the Web and use it for publishing the GISgenerated navigational maps on the Web while maintaining their links with the
underlying resource metadata base (these maps form the main part of
HealthCyberMap pilot service on the Web). It should be also possible to use the
same solution to publish other interactive GIS-driven maps on the Web, e.g., maps
of real-world health problems.
12. To develop a strategy for maintaining the currency of the generated hypermaps
and dealing with problems related to Web resource link persistence.
Evaluation Objectives
13. To identify the need and reasons for evaluating HealthCyberMap pilot service
interfaces on the Web, and investigate and develop/ refine suitable methods for
conducting this evaluation and analysing its results.
14. To conduct a small-scale formative evaluation of HealthCyberMap research pilot
using suitable methods (developed in the above objective), analyse the results, and
identify any useful lessons that could be learned and possible future directions that
could help improving HealthCyberMap if the service is further developed beyond
the current project plan and timeline.